Signal Presence Detectors?

[Don Tillman]

[A week ago I asked about ideas or schematics for signal presence
detectors.]

Martin and Juergen both suggested using a current steering technique
to drive the LEDs instead of simply switching the LEDs on.  This is a
*great* idea.  Otherwise, if I had a couple dozen LEDs triggering off
a common musical signal it would involve current glitches of 100mA or
more, sure to mess up a high fidelity audio signal somewhere.

(By "current steering" we mean using something like a current source
and a diff amp to drive the LED.  The same current is drawn whether
the LED is on or off.)

My current plans are to build the circuit like this:
  2-to-1 resistive divider to keep signals near the rails from being
  a problem.

  Single JFET source follower for some buffering and to isolate the
  input signal from the rest.

  Capacitively couple to a diff amp built from the 3 NPNs in a CA3096.
  The other input of the diff amp is biased at -0.050 volts; this is
  the voltage the signal needs to overcome to light the LED.

  A capacitor to average the DC value and a second PNP diff amp built
  from the remaining 2 PNPs in the CA3096 to drive the LED.

A quick breadboard circuit built with hand-matched discretes seems to
work well.  Thanks guys.

  -- Don

-- 
Don Tillman
Palo Alto, California, USA
don at till.com
http://www.till.com